Procedure for diphone synthesis and / or polifonemas from the actual frequency structure of the constituent phonemes.

Procedure for diphone synthesis and / or polifonemas from the actual frequency structure of the constituent phonemes. The present invention relates to the process for generating the acoustic signal synthesized voice sounds from a minimum prior information of the constituent phonemes obtained by segmentation of a previous recording; It said minimum information consists of the spectral envelope corresponding to a single period set in the stationary zone of the phoneme.